A hybrid approach based on FAHP and FIS for performance evaluation of employee Mohsen Sadegh Amalnick 1, Seyed Bahman Javadpour 2* Department of Industrial Engineering, College of Engineering, University of Tehran, Iran, Email: amalnick@ut.ac.ir. Department of Industrial Engineering, College of Engineering, University of Tehran, Iran, Abstract Human resources are the most important assets for every organization and their ways of behavior, operation, activities and functions could lead to the improvement the organization. The main aim of this study is to evaluate the performance of employees in an airline organization in Iran. The model of the study is tested on a sample of 14 employee of the mentioned organization in Iran using Fuzzy Analytic Hierarchy Process (FAHP) and Fuzzy Inference System (FIS). Keywords: Performance evaluation; FAHP; FIS; Fuzzy theory. 1. Introduction In today's world, organizational governance cannot be achieved solely by ingenuity and personal judgment, but decisions must be made based on scientific investigations, accurate and timely information and according to certain principles and procedures. Performance evaluation provides an appropriate context for both motivation and achievement of organizational goals by which one may be able to measure or assess the relationship between an individual s working hours and the amount of his/her work done. Evaluation has also been considered a highly effective tool for personal and professional enhancement of employees by which job distribution and delegation of authority may be accomplished based on the staff s merit. Kececi et al. (2015) offer a model for evaluating the performance of the naval officers, using Fuzzy Analytic Hierarchy Process (FAHP). They decide to evaluate the crew s performance by this model, due to their important role in maritime transport. Physical features of the workplace, job satisfaction, accessibility, competition, customs and working relationships are among the measures employed in their study. Chamoli (2015) presents a model by integrating the methods of Fuzzy Analytic Hierarchy Process (FAHP) and TOPSIS in a fuzzy environment and then applies it to evaluate the performance of air conditioning ducts. In order to find the optimal mode of air conditioning channels, several factors including the rate of air leakage, friction, and their efficiency and effectiveness, were investigated in this approach. Visalakshmi and Lakshmi (2015) offer a hybrid model based on DEMATEL and TOPSIS approaches in a fuzzy environment in order to evaluate economic performance of eco-friendly industries. Escrig-Olmedo et al (2015) propose a fuzzy TOPSIS to evaluate the performance of clothing industry with the aim of addressing the shareholders problems with the stability of the assets value and finally concluded by using specified measures. Hu et al (2015) provide a hybrid model using the fuzzy analytic network process (FANP) and DEMATEL to assess the qualitative performance of computer accessories suppliers in order to help the process of determining, comparing and ranking suppliers quality and finding the strengthening center of supply chain. Chen et al (2015) present a hybrid model of DEMATEL and fuzzy analytic network process (FANP) and employ the model to evaluate the performance of new product developments. 2. The main concern and the proposed approach for evaluation 57
In this the section, a new approach based on AHP and Fuzzy Inference System is presented for personnel evaluation. Due to hierarchical structure underlying the criteria and sub-criteria, AHP method was employed and for transferring the experts knowledge to the proposed approach, Fuzzy Inference System was used to create a decision support system. Fuzzy theory has also been used in order to import ambiguity and uncertainty in the problem. The proposed approach is implemented as follows: Step One: In this step, the effective criteria for evaluating the performance of staff should be selected. Step Two: In this step, we will determine the weight of each sub-criterion related to each adapted criteria using AHP. We use pairwise comparison matrix to specify the weight of sub-criteria and then determine the importance of the pairwise comparisons, using the table 1. Table 1: linguistic scales to determine the significance of paired comparisons Linguistic scales for difficulty Linguistic scales for importance Triangular fuzzy scale Triangular fuzzy reciprocal scale Just equal Just equal (1, 1, 1) (1, 1, 1) Equally difficult(ed) Equally importance(ei) (1/2, 1, 3/2) (2/3, 1, 2) Weakly more Weakly more difficult(wmd) importance(wmi) (1, 3/2, 2) (1/2, 2/3, 1) Strongly more Strongly more difficult(smd) importance(smi) (3/2, 2, 5/2) (2/5, 1/2, 2/3) Very strongly more Very Strongly more difficult(vsmd) importance(vsmi) (2, 5/2, 3) (1/3, 2/5, 1/2) Absolutely more Absolutely more difficult(amd) importance(ami) (5/2, 3, 7/2) (2/7, 1/3, 2/5) After the questionnaires were filled and pairwise comparison matrix extracted, each local weight factors will be obtained through a non-linear model which is given below. This model has been developed by Dağdeviren and Yuksel (2010): max λ s. t u ij m ij λw j + w i u ij w j 0 n m ij l ij λw j w i + l ij w j 0 w k = 1, w k > 0, k = 1,2,., n k=1 i = 1,2,., n 1, j = 2,3,, n, j > i In this nonlinear model (l, m, u) represents three triangular fuzzy numbers in the paired comparisons while w k indicates the weight of k th criterion. The optimum value of λ may be a positive or negative number. Positive values of λ imply that there is a compatibility in the pair comparison matrix which shows that comparisons have been properly judged. Negative values of λ, however, denote for incompatibility of the given matrix which. means that the experts should be asked to reconsider their judgments. This way, once the model is dissolved, we can obtain the local weight related to each criterion Step Three: In this step any of the employees score would be calculated for each criterion. To this end experts.are asked to apply linguistic terms using the table 2 58
Linguistic values for positive sub-factors Very weak Table 2: linguistic terms used in performance evaluation Linguistic values for Triangular fuzzy negative sub-factors numbers Very strong (0,0,0) The mean of fuzzy numbers 0 Weak Strong (0,0.167,0.333) 0.167 Weak-Mid Mid-Strong (0.167,0.333,0.5) 0.333 Mid Mid (0.333,0.5,0.667) 0.5 Mid-Strong Weak-Mid (0.5,0.667,) 0.667 Strong Weak (0.667,,1) Very strong Very weak (1,1,1) 1 3. Case study According to the steps mentioned in previous section, the related results will be discussed in this section. These steps are given below: Step One: As mentioned above, in this step performance evaluation criteria will be identified and extracted for airline employees. These criteria and sub-criteria are listed in table 3. Table 3: criteria and sub-criteria for performance evaluation Abbreviation Sub-criteria Criteria N1 Ability to reaching agreement with N2 Ability to communicate within the organization Innovation, N3 Ability to participate in teamwork Creativity and N4 Obligation to organization benefits Teamwork N5 Ability to creative thinking N6 Caring about company assets S1 Implementing client satisfaction plan Customer S2 Speed-up in client job satisfaction/ job S3 The way of answering to clients quality C1 Being patient C2 Dealing with criticism C3 Accepting the changes Personality and C4 Responsibility and dutifulness performance C5 Being on-time in work place factors C6 caring about appearance C7 Hygiene in work environment (implementing 5s) C8 Regarding safety, rules and regulation relating to organization Step two: in this step the sub-criteria presented in the previous step are weighted. To this end, we distribute the pairwise comparison questionnaires among the experts to complete them using the linguistic terms listed in table 1. Then, we will obtain the weights of the each sub-criteria using the Dağdeviren model (2010) discussed in the previous section and fuzzy pairwise comparison matrix. The weights of sub-criteria are listed in the following table: 59
Table 4: Weight of sub-criteria Weight Sub-criteria 0.342 Ability to reaching agreement with 0.031 Ability to communicate within the organization 0.162 Ability to participate in teamwork 0.162 Obligation to organization benefits 0.252 Ability to creative thinking 0.051 Caring about company assets 0.465 Implementing client satisfaction plan 0.211 Speed-up in client job 0.324 The way of answering to clients 0.182 Being patient 0.126 Dealing with criticism 0.128 Accepting the changes 0.188 Responsibility and dutifulness 0.014 Being on-time in work place 0.184 caring about appearance 0.154 Hygiene in work environment (implementing 5s) 0.024 Regarding safety, rules and regulation relating to organization Step three: in this step each of the employees will be evaluated for each of the criteria and then the scores obtained from evaluation will be calculated for everyone. Now, each employee s score should be calculated for each criterion. These scores which are listed in table 5, equal to the sum of sub-criteria weights multiplied by their numerical values. Table 5: the score obtained from performance evaluation for each criterion Employee Personality and Customer satisfaction/ Innovation, Creativity and performance factors job quality Teamwork Employee 1 0.4102 0.4325 0.3527 Employee 2 0.6254 0.5714 0.483 Employee 3 0.5377 0.5962 Employee 4 0.6838 0.69 0.672 Employee 5 0.4297 0.4628 0.4352 Employee 6 0.4694 0.6155 0.7126 Employee 7 0.7459 0.7205 0.6949 Employee 8 0.6458 0.8003 0.8659 Employee 9 0.5447 0.575 0.4988 Employee 10 0.6103 0.647 0.5918 Employee 11 0.6511 0.722 0.7029 Employee 12 0.5673 0.6101 Employee 13 0.6323 Employee 14 0.616 0.668 Step Four: In this step, the final score of each employee is calculated with the use of a decision support system based on fuzzy inference system. The development process of the fuzzy inference system is given below: In this step, we aim to create a decision support system based on fuzzy inference system. Therefore, in the first place, it is necessary to determine the input and output of the system and extract the fuzzy inference rules using the experts opinions. In this report, Score performance evaluation (Q) is the output while the inputs of the fuzzy inference system are: 1. Originality, creativity and teamwork (Q1) 2. Client satisfaction / quality of work (Q2) 3. Personality and functional factors (Q3) 60
The fuzzy inference rules are then extracted from the knowledge of experts and implemented in MATLAB. Using the data related to every employee (table 5) derived with the help of a support system, we calculate the score gained from performance evaluation for each employee. These results are given in Table 6. Employee Employee 1 Employee 2 Employee 3 Employee 4 Employee 5 Employee 6 Employee 7 Employee 8 Employee 9 Employee 10 Employee 11 Employee 12 Employee 13 Employee 14 Table 6: The final score of each employee Personality and Customer satisfaction/ Innovation, Creativity performance factors job quality and Teamwork 0.4102 0.4325 0.3527 0.6254 0.5714 0.483 0.5377 0.5962 0.6838 0.69 0.672 0.4297 0.4628 0.4352 0.4694 0.6155 0.7126 0.7459 0.7205 0.6949 0.6458 0.8003 0.8659 0.5447 0.575 0.4988 0.6103 0.647 0.5918 0.6511 0.722 0.7029 0.5673 0.6101 0.6323 0.616 0.668 Final score 0.414745 0.5482 0.680445 0.65333 0.474315 0.59864 0.71299 0.782625 0.565345 0.610355 0.68241 0.59828 0.584405 0.6287 These scores are numbers between 0 and 1. Decision makers, with the help of these scores would be able to classify their employees according to which they develop their goals and strategies based on human resources strategies, including motivational and promotional strategies, bonuses, fines and etc. 4. Conclusion This paper develops a performance evaluation model for airline. This model is then applied to a case study for the performance evaluation of 14 employee s airlines in Iran. We establish the procedures for identifying the most important criteria for assessment of employee of airline in Iran. The evaluation procedures consists of the following steps: (1) identify the evaluation criteria for airline; (2) assess the average importance of each criterion by Analytic Hierarchical Process over all the respondents. (3) Represent the performance assessment of employee for each criterion by fuzzy numbers. Finally, the final score of each employee are calculated by fuzzy inference system. References Visalakshmi, S., Lakshmi, P., Shama, M. S., & Vijayakumar, K. (2015). An integrated fuzzy DEMATEL-TOPSIS approach for financial performance evaluation of GREENEX industries. International Journal of Operational Research, 23(3), 340-362. Chamoli, S. (2015). Hybrid FAHP (fuzzy analytical hierarchy process)-ftopsis (fuzzy technique for order preference by similarity of an ideal solution) approach for performance evaluation of the V down perforated baffle roughened rectangular channel. Energy, 84, 432-442. Kececi, T., Bayraktar, D., & Arslan, O. (2015). A Ship Officer Performance Evaluation Model Using Fuzzy-AHP. Journal of Shipping and Ocean Engineering, 5, 26-43. Escrig-Olmedo, E., Fernández-Izquierdo, M. Á., Muñoz-Torres, M. J., & Rivera-Lirio, J. M. (2015). Fuzzy TOPSIS for an Integrative Sustainability Performance Assessment: A Proposal for Wearing Apparel Industry. In Scientific Methods for the Treatment of Uncertainty in Social Sciences (pp. 31-39). Springer International Publishing. Hu, H. Y., Chiu, S. I., Yen, T. M., & Cheng, C. C. (2015). Assessment of supplier quality performance of computer manufacturing industry by using ANP and DEMATEL. The TQM Journal, 27(1), 122-134. Chen, J. F., Hsieh, H. N., & Do, Q. H. (2015). Evaluating teaching performance based on fuzzy AHP and comprehensive evaluation approach. Applied Soft Computing, 28, 100-108. 61